Re: [PATCH V3] netfilter: netns nf_conntrack: per-netns net.netfilter.nf_conntrack_max sysctl

[Florian Westphal]

lvxiafei <xiafei_xupt@xxxxxxx> wrote:
> > in any case.
> >
> > Also:
> >
> > -       if (nf_conntrack_max && unlikely(ct_count > nf_conntrack_max)) {
> > +       if (net->ct.sysctl_max && unlikely(ct_count > min(nf_conntrack_max, net->ct.sysctl_max))) {
> >
> >
> > ... can't be right, this allows a 0 setting in the netns.
> > So, setting 0 in non-init-net must be disallowed.
> 
> Yes, setting 0 in non-init-net must be disallowed.
> 
> Should be used:
> unsigned int net_ct_sysctl_max = max(min(nf_conntrack_max, net->ct.sysctl_max), 0);
> if (nf_conntrack_max && unlikely(ct_count > net_ct_sysctl_max)) {

That would work.  Alternative, probably preferrable, is to do
something like this:

@@ -615,10 +615,10 @@ enum nf_ct_sysctl_index {
 static struct ctl_table nf_ct_sysctl_table[] = {
-               .proc_handler   = proc_dointvec,
+               .proc_handler   = proc_douintvec_minmax,
+               .extra1         = SYSCTL_ZERO, /* 0 == no limit */
        },
        [NF_SYSCTL_CT_COUNT] = {
                .procname       = "nf_conntrack_count",
@@ -1081,9 +1082,11 @@ static int nf_conntrack_standalone_init_sysctl(struct net *net)

        /* Don't allow non-init_net ns to alter global sysctls */
        if (!net_eq(&init_net, net)) {
                table[NF_SYSCTL_CT_EXPECT_MAX].mode = 0444;
                table[NF_SYSCTL_CT_BUCKETS].mode = 0444;
+
+               /* 0 means no limit, only allowed in init_net */
+               table[NF_SYSCTL_CT_MAX].extra1 = SYSCTL_ONE;
        }

That will make setting a 0 value illegal for non-init net case:

sysctl net.netfilter.nf_conntrack_max=0
sysctl: setting key "net.netfilter.nf_conntrack_max": Invalid argument

> min(nf_conntrack_max, net->ct.sysctl_max) is the upper limit of ct_count
> At the same time, when net->ct.sysctl_max == 0, the original intention is no limit,
> but it can be limited by nf_conntrack_max in different netns.

Sounds good to me.